All experiments in this thesis were carried out with adult Texel cross-bred ewes as experimental animals. The experiments were approved in advance by the regional ethics committee in Stockholm, Sweden and adhere to the European Union direc-tive 86/609/EEG and the European Council convention ETS 123.
THE SHEEP AS AN EXPERIMENTAL ANIMAL
Anaesthetics are intentionally used to reduce brain function. Although a range of different preparations exist they all achieve, through mechanisms not entirely known, interruption or disturbance of neuronal communication (Plourde 2007). It is thus not surprising that anaesthesia may modify the humoral, neural and cardiovas-cular responses to several different afferent stimuli (Dorward et al. 1985, Ninomiya et al. 1986, Weaver & Stein 1989). Therefore, studies investigating the central process-ing in cardiovascular regulation, due to its multiple integrative nature, should be preferably performed in conscious subjects.
In this regard the sheep is highly suited as a laboratory animal. These animals are domesticated and calm which facilitates experiments in the conscious state without the confounding influence of emotional stress. They are also favourably sized for multiple blood samplings with no major effect on the blood volume and chronic instrumentation without interfering with normal behaviour. Furthermore, haemody-namic parameters like MAP, CO, HR, several regional blood flows as well as circu-lating levels of AVP and ANG II largely corresponds to human values. However, all species have their anatomical and physiological peculiarities, more often than not associated with evolutionary adaptations to certain environmental demands. In rela-tion to the experiments included in this thesis some special features of the sheep de-serve attention.
In the sheep there is a common carotid artery on each side of the neck rather than separate internal and external carotid arteries. Arterial baro- and chemo-receptors are located at the junction of the carotid and occipital arteries. Sympathetic cardiac nerves enter the thorarcic sympathetic chain through spinal roots T1-T6 on the right side and T2-T5 on the left side and sometimes descend considerably before aiming for the heart. Neurons increasing heart rate are chiefly conferred to the right side (Waites 1957).
The kidneys are smaller in size compared to humans. They also receive slightly less blood in relation to CO. The right kidney is usually located against the liver in the right dorsal part of the abdomen while the left kidney is more mobile, the position being affected by the abdominal viscera.
Up to one fourth of the total erythrocyte mass can be stored in the spleen (Dooley et al. 1972), which is able to contract in response to sympathetic stimulation and add large quantities erythrocytes to the circulation. This feature is strongly activated dur-ing stress and blood loss and raise difficulties in estimatdur-ing plasma volume altera-tions by changes in haematocrit. In the undisturbed conscious state the haematocrit is around 30% while it falls to about 20% during isoflurane anaesthesia.
Also of importance in the haemorrhage situation is the reservoir of fluid residing in the forestomachs. Generally, it is the digestive system of the sheep that provides the greatest contrast to most of the other experimental animals. The forestomachs are of considerable size and contain fluid and electrolytes that may be recruited to the cir-culation during blood loss. Due to the great weight of the digestive system the blood volume related to body weight is lower compared to most mammals, including hu-mans. A sheep has approximately 60-65 ml blood per kg body weight (~ 4.5 litres for a 70 kg sheep) (Torrington et al. 1989).
Unlike monogastric animals the main source of energy are volatile fatty acids, which are produced by fermentation of cellulose and starches by bacteria in the rumen.
Consequently, the plasma glucose levels are lower than in man, reducing the ability to obtain acute hyperglycaemic hyperosmolality which may be of importance for the autotransfusion of fluid to the circulation after haemorrhage.
As herbivores, with less proteins/amino acids in the diet than omni- and carnivores the normal arterial pH of the sheep is normally slightly higher (between 7.45 and 7.50) than in humans. Hence, a pH within normal range for man may in the sheep be very low.
Figure 8. Creation of “arterial loops”. The carotid artery is first dissected free from surrounding tissue (A). Thereafter a rectan-gular area of skin, sufficiently large to enclose the artery, is vertically cut leaving the edges attached to the rest of the skin.
The artery is placed into the skin flap and the edges sutured to-gether (B). Finally, the skin un-derneath the loop is sutured (C).
SURGICAL PREPARATIONS
All surgical preparations were performed under isoflurane anaesthesia with the sheep orally intubated and mechanically ventilated. Postoperative treatment with analgesics and antibiotics was made routinely. None of the preparations caused any apparent discomfort to the sheep.
EXTERIORIZATION OF THE CAROTID ARTERIES
To achieve easy access to the arterial vascular compartment for pressure measure-ments and blood sampling the carotid arteries of the sheep were routinely enclosed into bilateral skin loops according to a technique described by Denton (Denton 1957). Great care was taken not to damage the vagus nerve in the process. Experi-mentation was not performed until the carotid loops had become satisfactory pliant for acute cannulation (approximately 2 weeks).
A
B
C
FLOW PROBES
The animals used in paper I-IV were supplied with ultrasonic flow-probes (Tran-sonic System Inc., New York, USA) for recordings of peripheral blood flow. The factory calibrated flow-probes are designed to accurately measure volume-flow, not only velocity. In papers II and III the flow-probes where placed around the femoral and the renal artery, while in papers I and IV an additional probe was used to meas-ure flow in the cranial mesenteric artery, corresponding to the superior mesenteric artery in man. The renal and mesenteric arteries were accessed via a paralumbar inci-sion on the left side. The descending aorta was dissected free and the appropriate vessels identified and supplied with the flow-probes as close as possible to the de-parture from the aorta. For the renal artery a probe size of 6 (reflector height 6.6 mm) was used and for the mesenteric the corresponding probe was 8 (reflector height 8.8 mm). The right femoral artery was uncovered with the sheep in the supine position and fitted with a probe size 6. All probe cables were then tunnelled subcu-taneously and fixed in place at a paralumbar position.
Figure 9. Placement of femoral artery flow-probe. An incision is made in the right groin of the sheep. The artery is located and the probe placed around it. The probe cables was then tun-nelled subcutaneously and fixed in place on the back of the sheep.
CEREBRAL CANNULAE
I.C.V.-cannulae
For access to the CSF compartment all animals were supplied with two permanently implanted stainless-steel guide tubes (o.d. 1.5 mm) placed with their tips 2-3 mm above each lateral cerebral ventricle. During the experiments, it was then easy to give infusions or take CSF-samples via an inner needle reaching beyond the tip of the guide tube. The procedure started with exposure of the skull bone. Thereafter two holes (diameter approximately 3 mm) were drilled 3-6 mm from the midline on ei-ther side. The tubes were lowered once at a time through the holes by the help of a micro manipulator. To ensure a correct placement, a slowly perfused (aCSF, 40 µL/min) inner needle extending past the guide cannula was connected to a pressure transducer. When the pressure dropped, indicating communication with the ventri-cle, the guide cannulae and supporting screws were secured to the skull with dental acrylic.
Figure 10. Example of a radiograph used in the placement of the PVN-cannulae. I.C.V.-injection of contrast medium visualized the cerebral ventricles. The location of the PVN was then esti-mated and the angle and depth of the cannulae calculated. The cannulae were advanced, one at a time, through drilled holes in the skull bone located 1 mm on either side of the mid-sagittal plane until their tips were situated 5 mm above the PVN. Dental acrylic secured the cannulae to the skull.
Abbreviations: 3v, third ventricle; I.C.V., intracerebroventricular; LV, lateral ventricle; MI, massa intermedia; PVN, nucleus paraventricularis;
PVN-cannulae
In paper V two additional guide tubes (o.d. 1.0 mm) were positioned in the brain 5 mm above the PVN bilaterally. The sheep were placed in a modified stereotactic apparatus and via injection of 0.6 ml iohexol contrast medium (Omnipaque, GE Healthcare Europe GmBh, Munich, Germany) the cerebral ventricles were visual-ized with radiographs. The anatomical location of the PVN was identified in a two dimensional sagittal plane and the angle and the depth of the tubes were calculated using basic trigonometry. Using a micromanipulator (David Kopf Instruments, Tu-junga, CA, USA) both tubes were lowered into the brain towards the PVN, but in different caudal to cranial angles, before they were fixed in place by dental acrylic. At the time of the experiment inner needles (o.d. 0.7 mm) were lowered through the guide tubes to reach the PVN.
All guide tubes were blocked with an obturator until the experiments were per-formed.
HAEMODYNAMIC MEASUREMENTS
Blood pressures were recorded from three different sources. First, a catheter (o.d. 1 mm) was inserted in one of the carotid arteries for measurement of arterial pressure.
From the pressure curve MAP and heart rate (HR) was derived. Second, a catheter (o.d. 1.2 mm) was inserted in the left jugular vein through which a guide wire was introduced into the vein. The catheter was retracted and a small incision made in the skin above the vein. By sliding on the wire another soft catheter was placed inside the jugular vein. A flow-directed, two lumen, pulmonary artery catheter was then slowly advanced via this introducer to the right side of the heart and further, until it was verified by pressure-guidance that it had reached a branch of the pulmonary ar-tery. The mean pulmonary artery pressure (MPA) was measured at the distal end of the catheter and the central venous pressure via a proximally placed port. Both catheters were connected to tubing containing heparinized saline that in turn were connected to pressure transducers (DPT-6003, PVB Medizin Technik, BMBH, Kirchseen, Germany). The voltage created by the saline pressing against a membrane inside the transducer was amplified using Biopac DAC100 (BIOPAC Systems; Go-leta, CA, USA). The transducers were calibrated before and after each experiment using a two-point calibration method with hydrostatic pressure corresponding to atmospheric pressure and 100 mmHg (arterial pressure) and atmospheric pressure and 25 mmHg (CVP and PAP).
The pulmonary artery catheter was also used for determination of CO via the ther-modilution technique (Ganz et al. 1971, Weisel et al. 1975). In papers I-III CO was computed every 30 to 60 seconds by automatic heating of the blood and
measure-ment of the rate of temperature drop along the catheter, whereas in papers IV and V this was supplemented by rapid I.V.of ice-cooled isotonic saline. The signal from the pulmonary artery catheter was fed into a Vigilance® Edwards Critical Care Monitor (Baxter Healthcare Corporation) for swift calculation of CO. The continuous ther-modilution technique has repeatedly been shown to give valid estimations of CO (Munro et al. 1994, Medin et al. 1998). However, during rapid changes in blood flow the automatic measurement have been reported to adapt slowly (Siegel et al. 1996).
In the haemorrhage experiments (papers I-IV) automatic CO measurements were successfully achieved until the decompensatory phase set in. Then the signal was sometimes lost for approximately five minutes until it recovered and reported drasti-cally lower CO-values. To confirm the major drop in CO manual measurements were added in paper IV.
Regional blood flows (femoral, renal and mesenteric) were measured by one or two dual channel flow-meters (T208, Transonic System Inc, New York, USA) connected to the implanted ultrasonic flow-probes.
All parameters were continuously recorded (except manual CO measurements) on a computer hard drive using a data acquisition system (MP150, BIOPAC Systems) with a sampling rate of 100 Hz (papers I-III) or 250 Hz (papers IV-V). A selected number of parameters were displayed on-line.
Intravascular catheterizations were performed under local anaesthesia at least 60 minutes prior to the start of the experiments to minimize the risk of possible stress effects.
EXPERIMENTAL PROTOCOLS
MILD HAEMORRHAGE
The haemorrhage procedure used in papers I-III and is illustrated in Fig. 11. Haem-orrhage was performed by aspiration of blood from the jugular vein at a constant rate (0.7 ml/min/kg) with the sheep standing unrestrained in an experimental cage (two experimental groups in paper I were anesthetized, lying in the prone position on an operating table). The blood was stored in heparinized bags for later retransfu-sion. Aspiration of blood continued until MAP dropped below 50 mmHg, i.e. the decompensatory phase set in. Prior to the start of haemorrhage, continuing throughout the experiment or stopped after a certain administered volume, an I.V. or
I.C.V. infusion was started. The difference in shed blood volume between infusions was taken as a measure of these treatments to postpone the decompensatory phase.
This protocol also permits evaluation of the haemodynamic responses to
haemor-rhage during the compensatory phase. However, comparisons between the infusions on the recovery after haemorrhage are rendered difficult if an effect is seen on the initiation of the decompensatory phase as this entail different degrees of hypovol-aemia when the recovery period starts.
The decompensatory phase was accompanied in the sheep with an increased respira-tory rate and sometimes the sheep would lie down. These reactions were not quanti-tatively or qualiquanti-tatively altered in relation to the volume of blood shed to induce hy-potension. The sheep never lost consciousness at the onset of hypotension, some-thing that is frequently seen in humans in the same situation.
MODERATE HAEMORRHAGE
This haemorrhage procedure was used in paper IV. It is modified from the method developed by Wiggers and Werle (Werle et al. 1942) and induces a state of reversible hypotensive and hypovolaemic shock. Here haemorrhage was performed as de-scribed above but the volume of blood loss was not guided by changes in MAP but instead was fixed at 25 ml/kg body weight in 25 minutes. To maintain hypotension an additional 10 ml/kg of blood was aspirated during 60 minutes. An I.C.V.-infusion was started after the first haemorrhage and continued until the end of the experi-ment. This protocol facilitates investigations of the central nervous component of different treatments given at a later stage, provided that the infused substance in-fused I.C.V.does not have any effects on the responses to haemorrhage per se.
The removed blood was always re-transfused, regardless of the haemorrhage proto-col used, and the sheep recovered rapidly after the experiments was stopped. As a result the sheep could be used as their own controls thereby eliminating the
inter-Figure 11. Schematic illustration of the “mild haemorrhage” experimental protocol.
Abbreviations: BW, body weight; I.C.V., intracerebroventricular; I.V., intravenous; MAP, mean arterial pressure.
IV- infusion (4 ml/kg during 60 min) ICV- infusion
30 min
Start haemorrhage
MAP<50 mmHg
Haemorrhage Recovery
0.7 ml/min/kg BW
Stop haemorrhage
60 min
individual variation in the statistical calculations. However, the haemorrhage experi-ments were separated with at least one (mild haemorrhage) or two (moderate haem-orrhage) weeks and the general condition of the sheep was always thoroughly evalu ated before and after each experiment.
REVERSIBLE INHIBITION OF NEURONAL ACTIVITY WITHIN THE PVN
This procedure was used in paper V. Lidocaine was utilized for inhibition of neu-ronal activity. This was done to investigate if inhibition of the neurotransmission within the PVN would influence the cardiovascular and renal effects of I.C.V. hyper-tonic (0.5 M) NaCl. Lidocaine acts on voltage-gated sodiumchannels by inhibiting the flow of sodium ions through the channel pore, thereby preventing the genera-tion and propagagenera-tion of depolarizagenera-tion. The effect is highly reversible, with the dura-tion being dose dependent. A disadvantage is that acdura-tions on the cell bodies can not be separated from actions on fibres of passage.
The temporary effect of lidocaine facilitated a cross-over design with a relatively short wash-out period for the investigation of the tonic contribution of the PVN to baseline haemodynamics as well as its significance for the cardiovascular and renal responses to increased CSF [Na+]. All sheep were randomized into two groups. In one group, the lidocaine intervention (A) was performed first and the control (B) after a two hour wash-out period. In the other group the order was reversed. In or-der to evaluate the tonic influence of the PVN on hemodynamic and renal parame-ters, the effect of lidocaine inhibition of the PVN was compared to control (no in-jection) and to injections of lidocaine placed outside the PVN. Thereafter lidocaine was injected again and the hypertonic NaCl infusion was started.
Figure 12. Schematic illustration of the “moderate haemorrhage” experimental protocol.
Abbreviations: I.C.V., intracerebroventricular.
25 ml/kg
Haemorrhage
85 min 180 min
30 min
Start haemorrhage
ICV-infusion
1 ml/kg/min 0.17 ml/kg/min
Stop haemorrhage Treatment 10 ml/kg
Recovery
BODY FLUID ANALYSES
Blood aspirated from the jugular vein was heparinized and used for immediate read-ing of haematocrit from 10 ml graded tubes after centrifugation (3000 RPM for 10 min, 1460 G). The plasma was taken for measurement of plasma osmolality (Auto &
Stat Om 6010 osmometer; Kagaku Co, Kyoto, Japan), [Na+] and [K+] (IL 943 flame photometer; Instrumentation Labs, Italy), protein concentration by refractometry (Atago Co, Kyoto, Japan). In paper V plasma as well as urine creatinine was deter-mined by the laboratory of the Karolinska University Hospital using the Jaffe method (Synchron LX, Beckman Instruments, Richmond, CA). Endogenous creatinine clearance has shown a good correlation to inulin clearance in sheep (Nawaz & Shah 1984) and was thus taken as an approximation of glomerular filtra-tion rate (GFR).
The carotid blood samples were used for immediate arterial blood gas analyses and sometimes for determinations of sodium and potassium concentrations (ion optodes using ion-selective ionophores) performed on an Opti Critical Care analyser (AVL, Roswell, Georgia, USA).
In paper V urine was collected in 60 min samples via a retention catheter inserted into the urinary bladder via the urethra. Urine volumes were measured and aliquots taken for determination of electrolytes, osmolality and creatinine concentration.
Lidocaine Lidocaine Lidocaine Lidocaine
0.5M NaCl ICV 0.5M NaCl ICV
A
B
0.5M NaCl ICV 0.5M NaCl ICV
Cont rol Cont rol
1 h
Cont rol
2 h 2 h
Cont rol
2 h
Wash-o ut B aseline
1 h
Figure 13. Schematic illustration of the experimental protocol utilized when the effect of PVN inhibition prior to I.C.V. hypertonic NaCl and per se was studied.
Abbreviations: I.C.V., intracerebroventricular.
CSF was sampled slowly and in small volumes (∼300 μl) via gentle siphoning from an I.C.V.-needle. The size of the sample only permitted analyses of [Na+] and [K+] with flame photometry.
VASOPRESSIN, RENIN ACTIVITY AND ANGIOTENSIN II
The analyses of these hormones were performed using radioimmunoassay (RIA), a method based on the competitive binding of radio-labelled and sample substances (in these cases peptides) to antibodies. Antibody specificity and sensitivity, as well as the stability of radio-labelling are crucial for the quality of the assay. Immediate cool-ing of plasma samples and deep-frozen storage are also of importance for reliable determinations.
The plasma renin activity was determined in papers I and III by using a commercial RIA kit (RIANEN Angiotensin [125I] RIA kit, Perkin Elmer). In this assay the amount of ANG I generated (per ml plasma and hour) during incubation at 37o C is determined by RIA, which gives an estimate of the renin activity. Renin activity as-says have generally been replaced by direct measurements of the renin concentra-tion. Since this particular assay was taken off the market, and to apply a somewhat less labour-intensive method, we switched to measurements of plasma ANG II con-centrations in papers IV and V.
For the determinations of AVP and ANG II, blood samples were collected in tubes with EDTA and Trasylol®, a protease inhibitor, and the separated plasma was fro-zen in two aliquots. The samples were later extracted with cold acetone and petro-leum benzene, followed by evaporation to dryness. These steps were taken to reduce degradation of peptides in the sample and eliminate possible disturbance by proteins and other substances on antibody binding to the peptide to be measured.
A commercial RIA kit (Peninsula laboratories Inc, Div. of Bachem, S-2012; RIK 7002) was used for plasma ANG II measurements. Three different assays were used for the plasma AVP measurements. In paper III a method developed in the labora-tory, and previously described in detail (Lishajko 1983) was used. The original assay used another extraction procedure (separation on Sephadex columns) than used here (see above). Due to possible deterioration of the antibodies (optimal degree of dilu-tion was declining) we changed to commercial RIA kits also for this peptide (paper I: Euria-Vasopressin, Euro-Diagnostica; papers IV and V: Peninsula Laboratories Inc, Div. of Bachem, S-2196; RIK 8103). All three assays gave similar basal plasma AVP levels in conscious sheep (1-3 pmol/L).
STATISTICAL ANALYSES
Values were presented as the arithmetic mean and an estimation of the variance in the measurement. Usually this was the standard deviation (S.D.) since it gives a better prediction of the variability than the sometimes used standard error of the mean (S.E.M.). S.E.M. is actually not a measure of variability but of uncertainty in the meas-urement. If effects were calculated (i.e. “a reduction in MAP increase by 8 mmHg compared to…”) the average discrepancy between groups were presented with the 95 % confidence interval of that difference. Although this occasionally makes the results section laborious to read it adds important information regarding the effect interval that contains the true population mean.
The sample sizes were generally small. Repeated measures in the same individual was performed when possible, to compensate for this. In case of non-significant results that were believed to be of interest, power-calculations were performed to put the negative findings in context.
As the dependent variables generally were followed over time the parametric method used was analysis of variance (ANOVA) with repeated measures. When the effects of different treatments (I.C.V. or I.V. infusions, PVN injections etc) were in-vestigated this factor was added to make a two- or three (anaesthesia in paper III) way ANOVA. If there was a significant interaction effect between time and treat-ment this was taken as an effect of the treattreat-ment. When the protocol consisted of several different treatments paired comparisons with Bonferroni correction was used to elucidate which treatment(s) was responsible for the interaction effect. If the P-value was equal to or below 0.05 the difference was considered significant.
The distribution of the data and variances were estimated from normal-distribution plots. In case of a non-normal distribution the data was first transformed, normally by taking the logarithm of the raw values. If the variable still did not pass the nor-mality test it was considered to have an unknown distribution. Consequently, it was statistically evaluated with nonparametric techniques such as Wilcoxon’s signed ranks test or Friedman’s ANOVA followed by Mann-Whitney’s test. (Altman 1991)